E38 SD idle fuelling inconsitent

[MadMaxHSV]

Despite the E40 being the neglected middle child, it's still holding more love from me than the E38.

Been battling SD tuning the VVE table on a friends car with a number of inconsistencies.

One of which is an inconsistent idle AFR. The car will start and idle badly in the 15-16's. Drive around a while and it will idle in the 12's. Stop the engine and restart...its back to 16. Blip the throttle and the AFR starts slowly falling towards 12.

This is all Open Loop SD. Idle at same RPM, same spark timing. Whats going on? Log of an example attached.
12145

All I can think is the spark timing does seem to creep down at the end maybe causing the richer mix, but only a couple of degrees difference?

[Taz]

Hello MadMaxHSV,

As a disclaimer, my comfort zone is the Gen III … still gaining experience with the Gen IV.

Took a look at the log … without knowing what WBO you are using - the actual AFR / Lambda is unknown, but the voltage change from AD4 could be interpreted as a fluctuating AFR or AFR drift.

Have you adjusted B1213 (Injector Flow Modifier IAT) to stabilize AFR with respect to intake air temperature ? (per Swingtan’s VVE Tuning Tutorial)

For OL tuning check the values of B0141 (OL Commanded Fuel in Gear), B0143 (OL Commanded Fuel in P/N), and B0146 (OL Fuel IVT Multiplier). Remember that B0141 and B0143 are in EQR.

The spark timing could be being removed by the ECM to lower the idle speed - not sure without seeing your tune.

Add these PIDs to future logs:

GM.IDLRPM (Desired Idle Speed)
E38.INVLVTMP_DMA (Intake Valve Temperature)

I have only done a few Gen IVs to date, but found the E38 Idle Tuning sticky by Hymey and Swingtan to be full of very applicable information - it is a little lengthy, but well worth the read. I condensed this into a few pages of hand written notes - which I refer to often.


Regards,
Taz

[MadMaxHSV]

Hi Taz,

Haven't dialled in the IAT modifier yet, will get this done to help clear the waters.

However in the log example the IAT's were not changing. Also regards the commanded fuelling OL, would this not show up in the AFR Commanded PID? (which is also constant at 14.7)

Idle spark would be the only thing on my current logs I can see as a possible variable. Will add desired idle to the PID list for the next logs.

Thanks

[Taz]

Out of town for the day - no access to software - any comments are from memory, so please check ...

Can't recall if the Commanded AFR PID is both OL & CL - or just CL. In the PIDs screen (F8) right click on the PID for more info. Tuned one in the heat of the summer this year and it took me a few attempts to get B1213 accurately represented - may or may not be contributory in your case - if memory serves your IAT was reasonably stable at 104 F.

Adding the intake valve temp PID will let you see which OL cells are being hit at idle.

Remember that OL fueling will be calculated based on VE (or VVE) and injector flow. Also, the ECM may cross reference VE with MAF data (if not running in SD mode).

I found the only way I could get the idle AFR the way I wanted was to dial in the VVE, and then modify the injector flow rate slightly in the idle kPa range - perhaps not textbook technique - but it works.

Our resident expert is Swingtan ... he would know the "correct" answer ...


Regards,
Taz

[swingtan]

Need some more info....


What's done to the car in terms of mods?
Post the current tune.
As Taz says, always log Inlet Valve Temp when working on the E38.
Info on the WB and what is the custom PID to convert the voltage to Lambda / AFR.


That'll do for starters....

Simon.

[MadMaxHSV]

1. Mods are stock LS2, TVS2300 supercharger with CAI. Siemens 60lb injectors (scaled 40%)
2. Current tune here 12151. OL SD with 2 bar MAP sensor. All airflow referencing tables have been scaled down 40% to match the injector scale. I did have to pull even more off the VVE to get it to idle (was far too rich initially). This may have been a source of error from the start.
3. WB is a NGK AFX. Calc PID is {EXT.AD4}*1.4+9

Sure i've made some school-boy error somewhere, with the VVE generations's shifting things out of place to boot.

[Taz]

With the NGK AFX use the following CALC PIDs (if plugged into AD4):

({EXT.AD4}*0.0960)+0.620 (this will display Lambda based on sensor voltage)
1/(({EXT.AD4}*0.0960)+0.620) (this will display EQR based on sensor voltage)
AFR is fuel dependent and therefore not consistent. The E38 does not have the GM EQR PID that the LS1-B PCMs feature - it does have AFR and Lambda though. With the E38 I work in Lambda. I generally prefer to work in EQR. I still log E38 AFR as a loose cross reference.

I only do NA (naturally aspirated) tunes - probably won't be of much assistance to you with this one.


Regards,
Taz

[MadMaxHSV]

Thanks Taz, i'm still in the NA portion of the VVE map if that helps ;)

[Taz]

OK ... again, I've only done a few E38s - and none were FI. It's night time on the the other side of the world, so Swingtan probably won't see this post for a while - to correct any misconceptions I have made.

Regarding idle ... the ECM controls air flow (and indirectly fuel flow) and spark timing to arrive at desire idle speed. Would highly recommend you try and absorb the information in the E38 Idle Tuning sticky mentioned in post #2 above.

You have "zeroed out" the Normal values in B1844 (Idle Integral Step size) and B1845 (Idle Proportional Step Size). This is done when tuning the idle to facilitate effectively adjusting B1652 (Max Idle Area 2 - stationary) - in isolation. After B1652 (and if necessary B1651 - Max Idle area 1 - non-stationary) are setup, B1844 & B1845 are then reactivated and adjusted. Followed by adjusting B1829 (Minimum Idle Airflow), and then finally the spark timing (B5128-B5131) is adjusted.

By setting B1844 & B1845 to zero, you essentially leave the ECM with only spark advance to attempt to control desired idle speed. This may be causing the engine to "load up" with fuel, if left idling for any length of time.

In a cammed / FI application I would also recommend zeroing out B5116 (Cold Engine Noise Reduction) and B5117 (Cold Engine Noise Reduction Multiplier).

Just thinking out loud ... this is not my area of expertise.


Regards,
Taz

[swingtan]

{B0179}: Predicted charge temp is set to "0" so the ECM will only use the IAT reading for the charge temp. If this was a dedicated race car, it's probably be OK. In a street car, you will end up with some problems with varying airflows. As the air speed drops, the charge temp will move toward ECT as it moves through the intake. With FI and intercoolers, the initial IAT will possibly be a bit higher than for NA, but the intercooler and in the case of the blower shorter runners, the over all effect is probably less. Where is the IAT sensor located? Is it measuring ambient air or the compressed air after the supercharger?
{B0180}: Setting of 1 is OK for SD tunes.
{B8101} / {B8103}: You VVE looks a bit rough and the values around the idle cells are low, even with the 40% scaling. It needs a bit of work.....


Over all though, I'm thinking that the issues is a mix of the VVE and the charge temp blending, but it's a bit hard to tell with the log data supplied. It looks like you are logging via the laptop, so can you load up the sample log attached to get the PID list, then clear the data (<CNTRL> +N), then add in "external analogue 4" for your WB.

Then turn the ignition on and start logging before you crank the engine over. Allow sufficient time for the WB to warm up before starting. Then just let the car idle with no throttle input. If possible do this for a cold start and a hot start. Then post the log back here.

12154

Simon.

[MadMaxHSV]

Hi Simon,

I'm sure your right. My only concern was that the idle was swinging either way so wasn't sure which way I needed to move the VVE.

The IAT is locate about 4 inches from the inlet valve after the charge cooler so should be pretty close to real IVT. I've been running my e40 SD with zero'd bias due to no MAF input needed on the e40 with no issues (IAT same place). But open to suggestions this may be effecting things. But certainly the stock values would be fairly wrong as the sensor is a long way from its original location in the MAF.

Agredd VVE is very rough, only a few road runs logged and still settling. Going to put it on a dyno to try and get some more settled numbers for BEN correction.

Can't get any more logs from this car for a few weeks as only working on it an afternoon at a time when time permits. Taking a while to settle the VVE on the road hence looking to use a dyno to speed this up and hopefully get more accurate results.

[swingtan]

OK, sounds like we will need to wait a bit for some more info.

[MadMaxHSV]

Yes sorry. Frustrating not being able to try things straight out.

Sure all thats happening is the idle is too lean and the ECM is somehow compensating swinging it rich. Likely richening up the idle will allow it to settle at a constant. Was curious to understand what methods the ECM has to interpret and affect fuelling when OL.

Will report back.

[joecar]

Take your time, no worries. :)